A Novel Psychological Decision-Making Approach for Healthcare Digital Transformation Benchmarking in ASEAN

Abstract

In recent years, digital transformation is seen as a mandatory and urgent requirement in the process of modernizing countries. The effectiveness of the digital transformation process in the field of public management directly affects the social life and operations of businesses. This study aims to paint an overall picture of the healthcare digital transformation of a rapidly growing region, the Association of Southeast Asian Nations (ASEAN), through a bounded rational multi-criteria assessment method. The novel proposed method is developed in light of the psychological behavior theories and the strengths of multi-criteria decision-making methods, which are based on distance computation and statistical parameters. Therefore, it can not only determine the weight of the criteria objectively through distance correlations, but also deeply describe the subjective psychology of the decision maker. In addition to theoretical contributions with a novel approach, the findings in the benchmarking process reveal important indicators and practical lessons from the digital transformation successes in ASEAN.

1. Introduction

Multiple criteria decision making (MCDM) is a field of study that deals with making decisions in complex and uncertain situations, where multiple objectives or criteria are involved. In such situations, it is often not possible or desirable to optimize for a single criterion, and instead, decisions need to be made by considering multiple criteria simultaneously [1]. MCDM provides a framework and tools for evaluating and ranking different alternatives based on multiple criteria, and for selecting the best option that satisfies the decision-maker’s preferences and objectives. MCDM is widely used in many fields, including business, engineering, environmental management, and public policy. MADM is a subset of MCDM that deals with situations where the decision-maker needs to choose the best alternative from a set of alternatives based on multiple attributes or criteria [2]. In MADM, the decision-maker needs to evaluate each alternative based on the same set of attributes or criteria, and then compare them to determine the best alternative [3]. Multi-criteria interactive decision-making (abbreviated as TODIM in Portuguese) is an MADM technique designed to address the ranking problem. TODIM contains aspects from both the aggregation approach and the outranking approach [4]. Fuzzy sets theory (FST) may lead to errors when translating accurate data into imprecise data. Hence, TODIM is a superior alternative to FST [5]. To determine objective criteria weighting, there are three main groups of methods that include subjective, objective, and integrated weighting [6,7]. Subjective methods rely on the opinions of experts or stakeholder groups [7,8,9], while objective methods use data from the decision-making matrix to evaluate criteria and their relationships [10,11,12]. Integrated methods combine multiple weighting techniques and normalize the resulting data [13,14,15]. One objective method called the CRiteria Importance Through Intercriteria Correlation (CRITIC) does not require input from the decision maker and is based on the difference in the structure of the judgment issue [16]. It is often used to determine the weight of qualities and does not require proof of attribute independence. Qualitative attributes can be converted into numerical variables for use in the computation process.

The key novelty of this study is the introduction of a Psychological Behavioral Distance-based Decision-making Method approach. The proposed method is inspired by principles of Prospect theory and Regret theory as well as other traditional MADM methods. At the beginning, the Regret theory is applied to convert the decision matrix to under-risk utility and regret utility. Then, the distance-based correlation computations are used to define the objective weight of criteria. Additionally, the dominance of alternatives is determined via the psychological behavioral distance-based utility matrices. In other words, the proposed method can not only objectively establish the weight of criteria through distance correlations, but also provide a profound understanding of the decision maker’s subjective psychology.

As the practical contribution, the proposed method is applied to evaluate the acceleration of digital transformation in the healthcare services of the Association of Southeast Asian Nations (ASEAN). Identifying and providing public services to vulnerable individuals exposes governments and humanitarian groups to several obstacles [17,18]. The goal of governments all over the world is to serve people with public services in a more effective manner by moving toward an open, bottom-up, agile, online, and integrated hierarchy of operations. Regardless of the current stage of digitalization, this will enable a new level of communication between citizens and their political representatives with many aspects such as education, non-communicable diseases, online transactional service, fiscal management, healthcare and citizens information. Furthermore, COVID-19 also significantly ramped up digitalization, which, among other things, now encourages people to imagine a variety of ways to utilize technology to optimize public service, thereby democratizing and improving the quality of life for millions of people. The value of public services is improved by cutting costs, boosting benefits, and extending access.

With the development of digital technology, it is now possible to offer services that are very helpful to the community on a large scale for a relatively low cost. Development organizations have applauded this movement, claiming that the use of digital technology enables governments to “overcome hurdles of location, poor physical connection, and inadequate administrative ability” and increases service acceptance, especially among underprivileged populations [19]. Since the 1990s, the phrase “digital transformation” has been used, though frequently under a different name, such as “e-government” [20], but there has lately been a renewed emphasis on digital media. Along with the technological revolution, such as in machine learning algorithms, there has been a shift in the narrative regarding user experience and interaction in service development and delivery. As a result of the many benefits of digitization, several nations around the world have used it to improve the performance of their public sectors, especially in healthcare landscapes [21].

Countries are rated based on their capacity to assure prompt delivery of services and the relevance of their respective factors. Furthermore, nations are classified as developed (first world), developing (second world), less developed, and undeveloped (third world) depending on these classifications. If a government can provide sufficient and satisfying public services to its residents, it acts as a stimulant to increase both personal and business production by building credibility [22]. Public service is the core aspect of government processes in all nations, regardless of their level of growth [23]. The ASEAN countries are developing nations that are trying to enhance the digitalization of public service and raise their residents’ living conditions. Health represents one of the country’s top priorities in providing public services. Currently, authorities in both developed and developing countries give the provision of public health services to their populations particular attention. Several e-health apps have been created by the commercial sector in Indonesia such as Klik Dokter and Halo Doc. Additionally, using big data to manage and reduce disaster risk in Philippines can be considered as a technological application. This country has been through natural disasters including typhoons, earthquakes, tsunamis, storm surges, mudslides, and volcanic activities. Due to inadequate preparedness for environmental disasters, hundreds of people are killed and injured annually, and billions of dollars are lost. Therefore, big data can be applied for climate change adaption and disaster risk reduction [24,25,26]. Therefore, not only has digitization-maintained pace with international tendencies in the period of the 4th industrial revolution, but it has also enabled the public sector to assist ASEAN countries in achieving their goal of becoming developed countries. Nevertheless, each country in Southeast Asia has a distinct political, social, and economic environment. As a result, the priority sectors for digitalization will vary based on these distinguishing criteria. The impact elements will serve as a comparison and reference tool for the effective digitization of public services in each nation. This paper will give a method for assessing the digitization of public service, emphasizing the prerequisites for utilizing the framework as well as policy proposals for the digital transformation journey and suggestions for the development of digital activities.

In the next section, we will examine relevant studies on the examination of digitalization. The steps involved in implementing the proposed method are set forth below in Section 3. In Section 4, a case study is provided. In Section 5, a commentary is given on the findings that were achieved, as well as a sensitivity analysis on the weight of each criterion. Section 6 contains a discussion of the closing comments as well as some directions for further study.

2. Literature Review

Over the last several years, many researchers have focused their attention on the digitalization sector [27]. Huisani et al. examines the effect of digitalization on both average and dispersed energy use [28]. Cross-sectional Augmented Autoregressive Distributed Lag (CS-ARDL) is used to assess both the long-run and short-run elasticity of digitalization. The findings indicate that digitization does reduce the overall and dispersed levels of energy use. Thus, if we want to reach our goals for energy sustainability, we should put more money into growing and improving our digital infrastructure. Numerous studies consider the global trend in digital applications for public service, and specifically, the ASEAN countries are one of the most interesting areas, concerning the sustainability of economic, social, and environmental pillars. This is because the ASEAN countries are home to some of the most rapidly developing economies in the world. Chee Wee Tan et al. created a criteria framework for connections in Singaporean firms to comprehend how organization–stakeholder interactions may be handled effectively to facilitate an appropriate reformation of business operations [29]. This study investigates in detail an e-government initiative since public agencies frequently adopt new information technology (IT) to simplify and innovate their obsolete processes. The researchers concentrated on the phenomenon of managing e-transformation, especially e-fill systems from 1992 to 2003. In addition, this research was to contribute to the academic and conceptual understanding of e-transformation by examining a variety of difficulties within the enterprise reformative procedures implicated in the implementation of an e-government curriculum. In Vietnam, Nguyen Hang Thanh et al. provided a case study on electronic customs about the digitalization of the public sector [30]. This research confirms that culture is a significant benefit, although finance, human resources and regulation may be obstacles. Currently, Vietnam is not constrained by technical advances. Additionally, administration rules and restrictions are a constraint. The paper’s findings provide authorities with crucial ideas for enhancing e-custom operations and aid company executives in formulating plans for adapting to the industrialization environment. Lhakard, Polwasit focuses on the supervisors of Thailand’s major public institutions as he examines Thailand’s digital transformation initiatives [31].

The government is pursuing digital transformation plans version 4.0 to enhance company efficiency, human well-being, and public service productivity. This study examined the difficulties in sustaining reform administration and responsibilities, regulation, accessibility, staff capacities, and performance evaluation, as well as the efforts of national agencies to solve these issues. The objectives of e-government include increasing the capability to provide improved public services, improving the quality of life for residents, boosting the competence and capability of the corporate sector, and strengthening community health and public safety. The whole ASEAN region will need to improve its internet platform, cultivate more talented individuals with superior technological knowledge, and implement a well-considered legislative authority [32]. Furthermore, for digitalization to work well in ASEAN countries, the public and private sectors need to work together and communicate with each other. From the literature review, it can be concluded that the digitization of public services has not been examined in ASEAN. Considering this, attempts have been made to evaluate this process in the current work. In fact, to the best of the authors’ understanding, this is the initial effort to apply MCDM analysis to assess the efficiency of this method. To fill the gap in research, important key criteria were first found through a review of the literature and then confirmed by talking to professionals. The analysis of decisions is a powerful approach to resolve problems involving various players, factors, and goals. MCDM issues typically consist of the following five elements: objective, selection designer’s choices, options, criteria, and results [33]. The Preference Ranking Organization Method for enrichment Evaluation (PROMETHEE) and the Elimination and Choice Expressing the REality (ELECTRE) are outstanding models that are widely utilized in energy planning. They are favored because they give a wide understanding of the published information and a practical perspective that raises all questions and suspicions. On the demand side, these strategies are the best ones to use when choosing where to put energy [34,35,36]. Moon Daeseop et al.’s report examines the profitability of the six transport routes between Korea and Europe [37]. Quantitative elements (total delivery time, travel distance and total transport cost) and qualitative components (transport system, safety, and knowledge) were chosen and weighted for the analytical standards. The Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) approach was then used on the criteria to rank the routes according to their effectiveness. Park Ji Yeong et al. discovered port intangible resources that may assist in the supply of high-quality port services and to offer a fuzzy TOPSIS method for solving the port selection issue with an emphasis on intangible resources [38]. Fuzzy TOPSIS is suitable for assisting with the decision-making process for unclear and uncertain issues, such as port selection in relation to intangible assets. In recent years, the TODIM method has been widely applied in decision making in many fields [39,40]. An expanded TODIM technique based on bidirectional projection is presented to handle the cooperative urgent decision-making issue in the setting of uncertain triangular fuzzy sets [41]. The unique method is utilized in a research study and compared to other classical techniques. When using a projection approach to calculate choice outcomes, two projection values will inevitably be identical. The combination of the expanded TODIM approach with the bidirectional projection method may provide more logical and accurate judgment outcomes. This is an indirect technique to illustrate the uniqueness and efficacy of the suggested approach and is another benefit of collective decision-making. On the other hand, the trend to apply psycho-behavioral theories to modify MADM methods is growing [42,43,44]. To represent these real-world decisions including emotions rather than optimum judgments (such as value maximization), traditional theories of decision behavior have been developed. Zhang Shitao et al. proposed a perceived utility value-based group consistency index (which takes into account the total continuity) and a group lack of consistency indicator (which reflects the total discrepancy) for pairwise standings of options based on regret theory and a scientific hypothesis multifaceted personal taste order specified by the decision makers for an air-fighter selection method [45]. Li men et al. proposed a Regret theory-based stochastic user equilibrium model of route selection by incorporating a regret aversion parameter to represent the regret level of passengers. The numerical findings collected from three example networks indicate that the regret aversion attitude does influence the route selection behavior of passengers and that the proposed model can more accurately describe the route selection behavior than current models [46]. Moreover, healthcare digitalization was processed in European Union (EU) member states. The MCDM technique can be used to assess the implementation of countries. In 2020, an evaluation for the two eHealth domains was stated. The objective is to assess the eHealth deployment in the EU Member States by synthesizing current composite indicators into an evaluation model based on multiple-criteria decision-making approaches [47]. The assessment is conducted using the TOPSIS, the Weighted Sum Approach (WSA), and the Multi-criteria Analysis of Preferences by means of Pair Actions and Criteria comparisons (MAPPAC) techniques. The outcome of the study is the development of an assessment methodology for eHealth that allows worldwide benchmarking of eHealth indicators and ranking-based ratings of eHealth deployment in European Union member states. Mohammadi Dehcheshmeh et al. suggested applying the analytical hierarchy procedure to establish the significance coefficients of the criterion and the multiple attribute decision-making technique [48]. The effectiveness of the e-government security strategy was evaluated using a new fuzzy MCDM that included pessimistic, normal, and optimistic perspectives on objects to support the policymaker’s decision [49]. The hurdles to the formation of such partnerships in the economic, infrastructural, legal, and social sectors of Iran’s transportation sector were explored. Using the Analytic Hierarchy Process (AHP), TOPSIS, and the Simple Additive Weighting (SAW) techniques of MCDM, the obstacles were then ranked according to their perceived importance. According to the findings of three methodologies, the primary impediment to the establishment of public–private collaborations is the aspect of limited economic markets and finance supply [50]. Mukhametzyanov et al. pointed out which approach is chosen can have a big effect on how decisions are made [51]. They also underlined the need to use many methodologies in the decision-making process in order to reach a long-term and excellent conclusion. Therefore, MCDM is initially proposed in this research, which is identified from the literature review and verified by consultation with professionals. Existing research on emergency decision making has significantly contributed to complex and uncertain decision environments, as shown by the evaluation of the relevant literature. On the one hand, several fuzzy theories have been used to represent and describe the fuzziness and ambiguity of many judgments gathered from decision makers. Although these fuzzy approaches have certain desirable qualities for managing ambiguous decision-making data, there are still scenarios in which complicated evaluations cannot be described correctly and random errors in the process cannot be effectively addressed.

From the review of the different methods, it can be seen that there is a tendency to combine MCDM methods for the process of determining the weight of the criteria and prioritizing the alternatives. In addition, fuzzy theory is also increasingly popularly applied in pairwise comparisons or scoring through linguistic variables. However, the weight of the evaluation criteria can also be objectively determined based on the nature of the data using statistical tools and parameters. In addition to the objective weights of the criteria, psychological behavioral theories were also applied in previous studies to determine the most appropriate solution for a particular decision maker. However, the absence of an approach, which takes advantage of the strengths of these research directions, leaves a research gap. As described in

Table 1. Previous relevant studies.

No.	Author	Year	Method	Data	Criteria Weighting	Psychological Behavior Theories
1	L. Lai et al. [52]	2018	AHP and overall prospect values	Crisp and Expert judgment	Expert judgment	Prospect theory
2	W. Yunna et al. [53]	2018	AHP and the prospect values	Crisp and Expert judgment	Expert judgment	Prospect theory
3	S. K. Wen et al. [54]	2019	Extended Z-MABAC Method	Expert judgment	Expert judgment	Regret theory
4	Z. Jianghong et al. [55]	2020	maximizing deviation, TOPSIS methods and modified PROMETHEE method	Expert judgment	Expert judgment	Regret theory
5	Liu et al. [56]	2021	TODIM and ELECTRE II	Expert judgment	Expert judgment	-
6	Taddese et al. [57]	2021	AHP and VIKOR	Expert judgment	Expert judgment	-
7	Utama et al. [58]	2021	AHP and TOPSIS	Crisp and Expert judgment	Expert judgment	-
8	M. Enes et al. [59]	2021	SWARA and WASPAS	Expert judgment	Expert judgment	-
9	G. Jianwei at al. [60]	2021	Probabilistic language term sets (PLTSs) and regret theory	Expert judgment	Expert judgment	Regret theory
10	L. Zhengmin et al. [61]	2022	group best–worst method (GBWM) and the proposed regret theory-based EVAMIX (the evaluation of mixed data)	Crisp and Expert judgment	Expert judgment	Regret theory
This study	Wang et al.	2023	Psychological Behavioral Distance-based Decision-making Method approach	Crisp	Statistical based	Prospect theory and regret theory

, the novel method that this paper was developed is based on the motivation of filling that gap.

3. Methodology

3.1. Psychological Behavioral in Decision Making

Over the decades, Prospect theory and Regret theory are two psychological theories that have been applied in multiple criteria decision making (MCDM) [62,63]. Prospect theory, developed by Kahneman and Tversky [64], proposes that people make decisions based on the perceived value of potential gains and losses, rather than objective probabilities. This theory suggests that people are risk-averse when it comes to potential losses, but risk-seeking when it comes to potential gains. In MCDM, Prospect theory can be applied by considering the decision maker’s attitudes towards risk and the potential outcomes of each alternative. The prospect value function is described as Equation (1).

$$f\left(\mathit{\Delta }x\right)=\left\{\begin{array}{cc}{\left(\mathit{\Delta }x\right)}^{\alpha }& ,\forall \mathit{\Delta }x\ge 0;0<\alpha <1\\ -{\gamma \left(-\mathit{\Delta }x\right)}^{\beta }& ,\forall \mathit{\Delta }x<0;0<\beta <1\end{array}\right.$$

(1)

where $\mathit{\Delta }x$ expresses the loss value or gain value relative to a reference point. The parameters $\gamma , \alpha ,$ and $\beta$ represent the loss aversion, the gain risk attitude, and the loss risk attitude of decision makers, respectively. Loss aversion demonstrates a pervasive avoidance of decisions that may result in losses. Risk attitude is an important choice reaction to ambiguity, impacted by perception. A larger value of $\alpha$ describes a more important attitude toward an amount of gain. On the contrary, a larger value of $\beta$ describes a more important attitude for a loss amount.

In Regret theory, developed by Amos Tversky and Daniel Kahneman, decision makers are more likely to make decisions that minimize the potential for regret, rather than maximizing the potential for success. In MCDM, regret theory can be applied by incorporating the potential for regret into the criteria used to evaluate alternatives. For example, if a decision involves a trade-off between two options, the potential for regret associated with choosing each option can be taken into account to arrive at a more accurate evaluation of the alternatives. The Regret theory can be described as follows:

Definition 1.

The under-risk utility value obtained from alternative X, with choosing consequences $x$, can be calculated as Equation (2):

$${utility}^{risk}\left(x\right)=x^{\phi }, 0<\phi <1$$

(2)

The value of $\phi$ represents the risk aversion coefficient of the decision maker, with a larger value indicating a lower degree of risk aversion. Empirical evidence suggests a value of 0.88 for $\phi$ [64].

Definition 2.

Let $x_1$ and $x_2$ be consequences of choosing alternative $X_1$ and $X_2$. The regret–rejoice value of choosing alternative $X_1$ rather than $X_2$, with choosing consequences $x_1$ and $x_2$, is estimated as Equation (3):

$${utility}^{regret}\left(x_1,x_2\right)=1-e^{-\lambda \left({utility}^{risk}\left(x_1\right)-{utility}^{risk}\left(x_2\right)\right)}, \lambda >0$$

(3)

The value of $\lambda$ express the regret aversion coefficient of the decision maker. If the value of the regret aversion coefficient is small, the decision maker has a lower degree of regret aversion. The value of ${utility}^{regret}\left(x_1,x_2\right)$ signifies either regret if $utility\left(x_1\right)$ is less than or equal to $utility\left(x_2\right)$ or rejoice if the opposite is true. Based on experiments, a suggested value for λ is 0.3 [64].

Definition 3.

Consider alternative $X_i\left(i=1\dots n\right)$ with choosing consequences are $x_i\left(i=1\dots n\right)$. The overall utility value is obtained by alternative $X_i$ can be defined as Equations (4) and (5) [65]:

$$utility\left(x_i\right)={utility}^{risk}\left(x_i\right)+{utility}^{regret}\left(x_i,x^{\ast }\right)$$

(4)

$$where$$

$$x^{\ast }=\underset{1\le i\le n}{\mathit{max}} x_i and {utility}^{regret}\left(x_i,x^{\ast }\right)\le 0$$

(5)

3.2. Psychological Behavioral Distance-Based Decision-Making Method (PBDbM)

Inspired by the principles of prospect theory and regret theory as well as the advantages of statistical-based MADM methods, this study introduces the novel Psychological Behavioral Distance-based Decision-making Method. The advantage of the proposed method is to determine the weight of the criteria and prioritize the alternatives objectively based on statistical parameter calculations. However, the calculation results also consider the psychological behavior of the decision maker through which the most suitable solutions are determined. As shown in Figure 1, the proposed method consists of the following steps:

Step 1. Establish the decision matrix consisting of $I$ $(i=1,\dots ,I)$ alternatives and $(j=1,\dots ,J)$ criteria/indicators as shown in Equation (6). In which, $a_{ij}$ presents the evaluation score of the $i$th alternative according to the $j$th criterion.

$$A=\left[\begin{array}{cccc}{a}_{11}& a_{12}& \cdots & a_{1J}\\ a_{21}& a_{22}& \cdots & a_{2J}\\ ⋮& ⋮& \ddots & ⋮\\ a_{I1}& a_{I2}& \cdots & a_{IJ}\end{array}\right]$$

(6)

Step 2. Establish the normalized decision matrix according to Equation (7).

$$b_{ij}=\left\{\begin{array}{cc}\frac{a_{ij}-\underset{1\le i\le I}{\min } \left(a_{ij}\right)}{\underset{1\le i\le I}{\max } \left(a_{ij}\right)-\underset{1\le i\le I}{\min } \left(a_{ij}\right)}& , if jth is beneficial criterion\\ \frac{\underset{1\le i\le I}{\max } \left(a_{ij}\right)- a_{ij}}{\underset{1\le i\le I}{\max } \left(a_{ij}\right)-\underset{1\le i\le I}{\min } \left(a_{ij}\right)}& , if jth is non-beneficial criterion\end{array}\right.$$

(7)

Step 3. Applying regret theory, the under-risk utility matrix is determined by Equation (8). In which, $\phi$ represents the risk aversion coefficient of the decision maker.

$${utility}_{ij}^{risk}=b_{ij}^{\phi }$$

(8)

Step 4. According to Equation (3), the regret utility matrix is established by Equation (9). In which, $\lambda$ represents the regret aversion coefficient of the decision maker.

$${utility}_{ij}^{regret}=1-e^{-\lambda \left({utility}_{ij}^{risk}-\underset{1\le i\le I}{\max } \left({utility}_{ij}^{risk}\right)\right)}, \lambda >0$$

(9)

Step 5. According to Equation (4), the final utility matrix is established by Equation (10).

$$x_{ij}={utility}_{ij}^{risk}+{utility}_{ij}^{regret}$$

(10)

Step 6. The standard deviation of each criterion is calculated according to Equation (11).

$$s_j=\sqrt{\frac{{\left({\Sigma }_{i=1}^I{x}_{ij}-{\stackrel{-}{x}}_j\right)}^2}{I-1}} where {\stackrel{-}{x}}_j=\frac{\underset{i=1}{\overset{I}{\Sigma }}{x}_{ij}}{I}$$

(11)

Step 7. For each criterion, the Euclidean distance matrix between the alternatives’ final utility is constructed according to the Equations (12) and (13).

$$D_j=\left[\begin{array}{cccc}0& d_{12}^j& \cdots & d_{1I}^j\\ d_{21}^j& 0& \cdots & d_{2I}^j\\ ⋮& ⋮& \ddots & ⋮\\ d_{I1}^j& d_{I2}^j& \cdots & 0\end{array}\right] ,j=1,\dots ,J$$

(12)

$$where$$

$$d_{ik}=\begin{array}{cc}\left|x_{ij}-x_{kj}\right|& ,j=1,\dots ,J\end{array}, i=1,\dots ,I, k=1,\dots ,I, i\ne k$$

(13)

Step 8. For each Euclidean distance matrix, the double-centering process is performed. Accordingly, the row means, column means, and matrix means of the Euclidean distance matrix are determined. The row mean and column mean are deducted sequentially from each element. Then, the matrix mean is added to each element of the distance matrix. As the results, the double-centered matrices are constructed for each criterion as shown in Equation (14).

$$T_j=\left[\begin{array}{cccc}{t}_{11}^j& t_{12}^j& \cdots & t_{1I}^j\\ t_{21}^j& t_{22}^j& \cdots & t_{2I}^j\\ ⋮& ⋮& \ddots & ⋮\\ t_{I1}^j& t_{I2}^j& \cdots & t_{II}^j\end{array}\right] ,j=1,\dots ,J$$

(14)

Step 9. The distance covariance between $j$th criterion and $j^{’}$th criterion (${dCOV}_{jj’}$) is determined by the following sub-steps:

• Multiply double-centered matrices elementwise using the Hadamard product, also known as the Schur product or the entry wise product.
• Calculate the average value of the elementwise multiplication matrix.
• Calculate the square root of the average value.

Step 10. The distance variance of the $j$th criterion (${dVAR}_j={dCOV}_{jj}$) is determined similarly to step 9.

Step 11. The distance correlation between the $j$th criterion and $j^{’}$th criterion (${dCOR}_{jj’}$) is determined according to Equation (15).

$${dCOR}_{jj’}= \frac{{dCOV}_{jj}}{\sqrt{{dVAR}_j\times {dVAR}_{j’}}}$$

(15)

Step 12. The information content of the $j$th criterion ($I_j$) is determined according to Equation (16).

$$I_j=s_j\underset{j^{’}=1}{\overset{I}{\Sigma }}\left(1-{dCOR}_{j{j}^{’}}\right)$$

(16)

Step 13. The absolute weight of the $j$th criterion ($w_j$) is determined according to Equation (17).

$$w_j= \frac{I_j}{{\Sigma }_{j=1}^I{I}_j}$$

(17)

Step 14. According to the principle of reference dependence of the Prospect theory, the relative weight of $j$th criterion (${\widehat{w}}_j$) is estimated based on the absolute weight of reference criterion as Equation (18).

$${\widehat{w}}_j= \frac{w_j}{w_r} , where w_r=\underset{1\le i\le I}{\max } \left(w_j\right)$$

(18)

Step 15. The contribution of each $j$th criterion (${\mathsf{\omega }}_{ik}^j$) to the level of dominance $\left({\theta }_{ik}\right)$ of $i$th alternative over $k$th alternative is shown in Equations (19) and (20). Accordingly, if $x_{ij}>x_{kj}$, the value of ${\mathsf{\omega }}_{ik}^j$ will be a gain to ${\theta }_{ik}$. if $x_{ij}<x_{kj}$, it will be a loss to ${\theta }_{ik}$. Otherwise, the $j$th criterion has no contribution to ${\theta }_{ik}$. Besides, according to Prospect theory, the loss attenuation coefficient $\mu$ is used to express the loss attenuation of the decision makers.

$${\mathsf{\omega }}_{ik}^j=\left\{\begin{array}{cc}\sqrt{\frac{{\widehat{w}}_j(s_{ic}^{\ast }-s_{jc}^{\ast })}{{\Sigma }_{j=1}^J{\widehat{w}}_j}}& ,if x_{ij}>x_{kj}\\ 0& ,if x_{ij}=x_{kj}\\ -\frac{1}{\mu }\sqrt{\frac{\left(s_{jc}^{\ast }-s_{ic}^{\ast }\right){\Sigma }_{j=1}^J{\widehat{w}}_j}{{\widehat{w}}_j}}& ,if{x}_{ij}<x_{kj}\end{array}\right.$$

(19)

$${\theta }_{ik}=\underset{j=1}{\overset{J}{\Sigma }}{\mathsf{\omega }}_{ik}^j$$

(20)

Step 16. The total dominance of the $i$th alternative is calculated according to Equation (21).

$${\theta }_i^{\ast }=\underset{k=1, k\ne i}{\overset{I}{\Sigma }}{\theta }_{ik}$$

(21)

Step 17. The overall evaluation score (${\tau }_i$) of the $i$th alternative is determined as Equation (22). The higher the value of ${\tau }_i$, the better the alternative.

$${\tau }_i=\frac{{\theta }_i^{\ast }- {min}_{i=1}^I{\theta }_i^{\ast }}{{max}_{i=1}^I{\theta }_i^{\ast }- {min}_{i=1}^I{\theta }_i^{\ast }}$$

(22)

4. Numerical Results

4.1. Indicator Identification and Data Collection

According to previous research and a discussion with experts, indicators for the digitalization of healthcare parameters and their descriptions are established.

The United Nations’ (UN) E-Participation Index (EPI) has evaluated global e-participation activities since 2003. The E-Participation Index (EPI) is calculated and applied to evaluate e-participation projects in nations throughout the globe according to a complement to the UN E-Government Survey [66]. Recognizing e-participation begins with its underlying distribution. It begins, as a prerequisite, with the instructive stage, in which the state offers its elements with basic knowledge, followed by the two components, in which individuals are requested to provide insight to authorities, and finally, “the collaboration alternative”, in which residents become the pioneers by directing the strategy procedure. This latter structure strongly matches the UN E-Participation Framework’s three-tiered layout. The EPI has been a multidimensional framework consisting of three fundamental parts: e-consultation, e-information and e-decision-making [67]. The objective of this indicator is not to advocate a particular approach, but rather to provide perspective into how other nations are using internet resources to promote communication between the state and its citizens, as well as with the participants, to the advantage of everyone. In this study, EPI is known as Indicator 1 (C1).

Online service index (OSI), Human capital index (HCI) and Telecommunication infrastructure index (TII) are the Indicators C2, C4 and C5, respectively. These indicators are elements of the E-Government Development Index (EGDI), which is the moderate of three standard ratings of the three most important things about e-government. The E-Government Development Index illustrates the level of e-government development among member governments. In addition to an evaluation of a nation’s website-building trends, the E-Government Development Index combines key factors, such as facilities and parental education, to demonstrate how a country uses digital technology to enhance the inclusion and accessibility of its citizens. The EGDI is a comprehensive index of three crucial facets of e-government: telecommunication connection, online service supply and human capability [68]. Measuring nationwide and regional metrics of e-government growth may aid authorities in gaining a better understanding of the real state of e-government. EGDI evaluates the growth of e-government in many nations [69,70]. Consequently, the government may predict future trends and overcome impediments by analyzing its own growth trajectory. Government effectiveness covers views of the performance of public services, the competence of the public sector and its independence from constraints, the efficiency of policy creation and execution, and the trustworthiness of the company’s ability to achieve such objectives, which is as known as indicator C3. The percentile rank shows where the country stands among all the states represented by the extracellular matrix. The lowest rank is 0, and the highest rank is 100. Adjustments have been made to the percentile rankings to account for variations in the makeup of the Worldwide Governance indicator’s included nations over time. Therefore, indicator 4 is known as government effectiveness [71]. Universal Health Coverage (UHC) refers to making certain that all individuals have access to the medical treatments they need—without experiencing economic difficulties—and is essential for enhancing the nation’s well-being. UHC is also a human resource investment and a fundamental promoter of equitable and long-term financial growth and advancement. Besides, tracer measures comprise reproductive, maternal, neonatal, and health outcomes, viral illnesses, medical complications, provision of service, and accessibility. The measure of coverage for critical medical services is displayed on a 0–100 scale [72]. In this study, UHC is considered as indicator C6. Indicator C7 is secure Internet servers, which is the number of selected, clearly and openly TLS/SSL certificates discovered by the Netcraft Secure Server Survey [73]. In addition, population is an essential metric because it influences the amount of time required to distribute technology and information technology to a whole country’s population. Consequently, indicator C8 as known as population.

A case study is applied for the proposed novel approach to evaluation digital transformation process among ASEAN Countries according to the above eight indicators. Furthermore, actual data are acquired for indices ranging from C1 to C5. The Grey Prediction Model, also known as GM (1,1) has been used to provide forecast values of indicators C6, C7, and C8 in the year 2022 based on historical data. Because of the popularity of the GM (1,1), descriptions of this model are omitted in this article.

Table 2. Decision matrix.

Country	Code	C1 (0–1 Scale)	C2 (0–1 Scale)	C3 (%)	C4 (0–1 Scale)	C5 (0–1 Scale)	C6 (%)	C7 (No. of TLS/SSL)	C8 (People)
Brunei	BN	0.4773	0.5871	91.3462	0.8372	0.7567	81.2944	14,703	517,740
Cambodia	KH	0.2841	0.4181	36.0577	0.5605	0.538	69.137	651,189	298,238,245
Indonesia	ID	0.7159	0.7644	65.3846	0.6397	0.7438	69.3884	4521	17,883,678
Lao PDR	LA	0.2614	0.3005	30.2885	0.282	0.5468	56.5393	1054	8,009,566
Malaysia	MY	0.6818	0.763	81.25	0.7945	0.7645	56.5393	1016	58,328,992
Myanmar	MM	0.3068	0.3073	8.65385	0.6082	0.5829	80.8361	308,704	37,242,523
Philippines	PH	0.4886	0.6303	57.6923	0.5638	0.7629	62.3952	13,303	122,056,425
Singapore	SG	0.9773	0.962	100	0.8758	0.9021	93.4973	796,767	6,268,585
Thailand	TH	0.7841	0.7763	60.5769	0.7338	0.7879	90.173	243,223	80,502,792
Vietnam	VN	0.5341	0.6484	62.0192	0.6973	0.6903	80.7708	438,126	106,757,032

is a tabular presentation of the data that serves as a summary, where all of indicators are considered as beneficial indicators.

4.2. Indicatior Weighting

Table 3. Normalized decision matrix.

Country	C1	C2	C3	C4	C5	C6	C7	C8
Brunei Darussalam	0.302	0.433	0.905	0.935	0.601	0.670	0.017	0.000
Cambodia	0.032	0.178	0.300	0.469	0.000	0.341	0.817	1.000
Indonesia	0.635	0.701	0.621	0.602	0.565	0.348	0.004	0.058
Lao PDR	0.000	0.000	0.237	0.000	0.024	0.000	0.000	0.025
Malaysia	0.587	0.699	0.795	0.863	0.622	0.000	0.000	0.194
Myanmar	0.063	0.010	0.000	0.549	0.123	0.657	0.387	0.123
Philippines	0.317	0.499	0.537	0.475	0.618	0.158	0.015	0.408
Singapore	1.000	1.000	1.000	1.000	1.000	1.000	1.000	0.019
Thailand	0.730	0.719	0.568	0.761	0.686	0.910	0.304	0.269
Vietnam	0.381	0.526	0.584	0.699	0.418	0.656	0.549	0.357

presents the normalized decision matrix, which is constructed based on the ten alternatives’ indicators and Equation (7). The under-risk utility matrix is derived in

Table 4. Under-risk utility matrix with risk coefficient = 0.88.

Country	C1	C2	C3	C4	C5	C6	C7	C8
Brunei Darussalam	0.348	0.479	0.916	0.943	0.639	0.703	0.028	0.000
Cambodia	0.048	0.219	0.347	0.514	0.000	0.388	0.837	1.000
Indonesia	0.670	0.732	0.658	0.640	0.605	0.395	0.008	0.082
Lao PDR	0.000	0.000	0.282	0.000	0.038	0.000	0.000	0.039
Malaysia	0.626	0.730	0.817	0.878	0.659	0.000	0.000	0.236
Myanmar	0.088	0.018	0.000	0.590	0.159	0.691	0.433	0.159
Philippines	0.364	0.542	0.578	0.519	0.654	0.198	0.025	0.455
Singapore	1.000	1.000	1.000	1.000	1.000	1.000	1.000	0.031
Thailand	0.758	0.748	0.608	0.786	0.718	0.920	0.351	0.315
Vietnam	0.428	0.568	0.623	0.730	0.464	0.690	0.590	0.404

by utilizing Equation (8). As indicated in

Table 5. Regret Utility matrix with Regret coefficient = 0.3.

Country	C1	C2	C3	C4	C5	C6	C7	C8
Brunei Darussalam	−0.216	−0.169	−0.025	−0.017	−0.115	−0.093	−0.339	−0.350
Cambodia	−0.331	−0.264	−0.217	−0.157	−0.350	−0.202	−0.050	0.000
Indonesia	−0.104	−0.084	−0.108	−0.114	−0.126	−0.199	−0.346	−0.317
Lao PDR	−0.350	−0.350	−0.241	−0.350	−0.335	−0.350	−0.350	−0.334
Malaysia	−0.119	−0.084	−0.056	−0.037	−0.108	−0.350	−0.350	−0.257
Myanmar	−0.315	−0.343	−0.350	−0.131	−0.287	−0.097	−0.185	−0.287
Philippines	−0.210	−0.147	−0.135	−0.155	−0.109	−0.272	−0.340	−0.178
Singapore	0.000	0.000	0.000	0.000	0.000	0.000	0.000	−0.337
Thailand	−0.075	−0.078	−0.125	−0.066	−0.088	−0.024	−0.215	−0.228
Vietnam	−0.187	−0.138	−0.120	−0.084	−0.174	−0.098	−0.131	−0.196

, the regret utility matrix is then constructed using Equation (9). According to Equation (10),

Table 6. Final utility matrix.

Country	C1	C2	C3	C4	C5	C6	C7	C8
Brunei Darussalam	0.132	0.310	0.891	0.925	0.524	0.610	−0.311	−0.350
Cambodia	−0.283	−0.045	0.130	0.357	−0.350	0.186	0.787	1.000
Indonesia	0.567	0.648	0.549	0.526	0.480	0.196	−0.338	−0.235
Lao PDR	−0.350	−0.350	0.041	−0.350	−0.297	−0.350	−0.350	−0.295
Malaysia	0.507	0.645	0.761	0.841	0.551	−0.350	−0.350	−0.021
Myanmar	−0.226	−0.325	−0.350	0.459	−0.129	0.594	0.248	−0.129
Philippines	0.154	0.395	0.444	0.364	0.545	−0.074	−0.314	0.277
Singapore	1.000	1.000	1.000	1.000	1.000	1.000	1.000	−0.306
Thailand	0.683	0.670	0.484	0.720	0.630	0.896	0.136	0.086
Vietnam	0.240	0.430	0.503	0.646	0.290	0.592	0.459	0.208
Std	0.448	0.449	0.410	0.388	0.442	0.485	0.513	0.408

illustrates the final utility matrix. The standard deviation is determined for all eight indicators using Equation (11). Then, based on Equations (12) and (13), the Euclidean distance matrix between the alternatives’ total utilities is estimated.

Table A1. Euclidean distance matrix for criterion C1.

Country	BN	KH	ID	LA	MY	MM	PH	SG	TH	VN	Row Mean
BN	0.000	0.415	0.434	0.482	0.375	0.359	0.022	0.868	0.551	0.108	0.361
KH	0.415	0.000	0.849	0.067	0.790	0.056	0.437	1.283	0.966	0.523	0.539
ID	0.434	0.849	0.000	0.916	0.059	0.793	0.412	0.412	0.116	0.326	0.432
LA	0.482	0.067	0.916	0.000	0.857	0.124	0.504	1.350	1.033	0.590	0.592
MY	0.375	0.790	0.059	0.857	0.000	0.733	0.353	0.493	0.176	0.267	0.410
MM	0.359	0.056	0.793	0.124	0.733	0.000	0.380	1.226	0.909	0.467	0.505
PH	0.022	0.437	0.412	0.504	0.353	0.380	0.000	0.846	0.529	0.086	0.357
SG	0.868	1.283	0.412	1.350	0.493	1.226	0.846	0.000	0.317	0.760	0.755
TH	0.551	0.966	0.116	1.033	0.176	0.909	0.529	0.317	0.000	0.443	0.504
VN	0.108	0.523	0.326	0.590	0.267	0.467	0.086	0.760	0.443	0.000	0.357
Column mean	0.361	0.539	0.432	0.592	0.410	0.505	0.357	0.755	0.504	0.357	0.481

is an example of an indicator C1, Euclidean Distance matrix. In addition, recalculating using the remaining indications based on Equation (14), the mean of the additional matrix is calculated as shown in

Table A2. The double-centering matrix of alternative according to indicator C1.

Country	BN	KH	ID	LA	MY	MM	PH	SG	TH	VN
BN	−0.241	−0.004	0.122	0.010	0.085	−0.026	−0.215	0.232	0.167	−0.129
KH	−0.004	−0.596	0.360	−0.582	0.322	−0.506	0.022	0.470	0.404	0.109
ID	0.122	0.360	−0.383	0.373	−0.302	0.337	0.105	−0.294	−0.338	0.018
LA	0.010	−0.582	0.373	−0.703	0.336	−0.492	0.036	0.483	0.418	0.122
MY	0.085	0.322	−0.302	0.336	−0.339	0.300	0.067	−0.192	−0.257	−0.019
MM	−0.026	−0.506	0.337	−0.492	0.300	−0.528	0.000	0.447	0.382	0.086
PH	−0.215	0.022	0.105	0.036	0.067	0.000	−0.233	0.215	0.149	−0.146
SG	0.232	0.470	−0.294	0.483	−0.192	0.447	0.215	−1.030	−0.461	0.128
TH	0.167	0.404	−0.338	0.418	−0.257	0.382	0.149	−0.461	−0.527	0.063
VN	−0.129	0.109	0.018	0.122	−0.019	0.086	−0.146	0.128	0.063	−0.233

. Next, the distance variance of indicators is calculated as shown in

Table A3. The distance variance matrix of alternatives according to indicator C1.

Country	BN	KH	ID	LA	MY	MM	PH	SG	TH	VN
BN	0.058	0.000	0.015	0.000	0.007	0.001	0.046	0.054	0.028	0.017
KH	0.000	0.355	0.130	0.339	0.104	0.256	0.001	0.221	0.164	0.012
ID	0.015	0.130	0.146	0.139	0.091	0.114	0.011	0.086	0.114	0.000
LA	0.000	0.339	0.139	0.495	0.113	0.242	0.001	0.234	0.175	0.015
MY	0.007	0.104	0.091	0.113	0.115	0.090	0.005	0.037	0.066	0.000
MM	0.001	0.256	0.114	0.242	0.090	0.279	0.000	0.200	0.146	0.007
PH	0.046	0.001	0.011	0.001	0.005	0.000	0.054	0.046	0.022	0.021
SG	0.054	0.221	0.086	0.234	0.037	0.200	0.046	1.060	0.213	0.017
TH	0.028	0.164	0.114	0.175	0.066	0.146	0.022	0.213	0.277	0.004
VN	0.017	0.012	0.000	0.015	0.000	0.007	0.021	0.017	0.004	0.054

. The computation is then repeated with the remaining indicators. The distance covariance between the C1 and C2 indicators is shown in

Table A4. The distance covariance of alternatives according to indicator between C1 and C2.

Country	BN	KH	ID	LA	MY	MM	PH	SG	TH	VN
BN	0.060	0.000	0.008	0.001	0.005	−0.002	0.031	0.038	0.012	0.011
KH	0.000	0.317	0.099	0.242	0.088	0.213	0.002	0.176	0.115	0.014
ID	0.008	0.099	0.116	0.147	0.090	0.131	−0.001	0.060	0.099	0.000
LA	0.001	0.242	0.147	0.638	0.131	0.425	0.007	0.238	0.168	0.030
MY	0.005	0.088	0.090	0.131	0.102	0.116	0.000	0.038	0.075	0.000
MM	−0.002	0.213	0.131	0.425	0.116	0.458	0.000	0.218	0.151	0.020
PH	0.031	0.002	−0.001	0.007	0.000	0.000	0.050	0.020	0.001	0.023
SG	0.038	0.176	0.060	0.238	0.038	0.218	0.020	0.833	0.110	−0.020
TH	0.012	0.115	0.099	0.168	0.075	0.151	0.001	0.110	0.173	0.000
VN	0.011	0.014	0.000	0.030	0.000	0.020	0.023	−0.020	0.000	0.039

. The computation is then repeated with the remaining indicators. The $dCOR$ matrix is shown in

Table A5. The distance correlation matrix.

Indicator	C1	C2	C3	C4	C5	C6	C7	C8
C1	1.000	0.965	0.818	0.725	0.906	0.484	0.404	0.446
C2	0.965	1.000	0.874	0.734	0.928	0.424	0.376	0.444
C3	0.818	0.874	1.000	0.823	0.885	0.398	0.385	0.525
C4	0.725	0.734	0.823	1.000	0.750	0.630	0.365	0.514
C5	0.906	0.928	0.885	0.750	1.000	0.466	0.471	0.463
C6	0.484	0.424	0.398	0.630	0.466	1.000	0.637	0.341
C7	0.404	0.376	0.385	0.365	0.471	0.637	1.000	0.515
C8	0.446	0.444	0.525	0.514	0.463	0.341	0.515	1.000

. The weight of indicators is calculated according to Equation (17) and is shown in Figure 2. Consequently, secure internet servers (C7), UHC service coverage index (C6), and Population (C8) are the three most significant indicators, with respective values of 0.195, 0.173, and 0.151. The weight of the Government Effectiveness (C3) and Human capital Index factors (C5) with the least significance is 0.093. Meanwhile, there is no substantial variation in weight among the other indicators.

Consequently, the quality of an economy’s infrastructure, especially its electricity and connectivity, is a crucial factor for both local and international investors when making investment choices. Information and communication technologies are widely acknowledged as crucial instruments for advancement, assisting in global cooperation and promoting public sector performance, productivity, and accountability. Moreover, the responsiveness of a nation’s health system to necessary services is crucial and needed. This indicates the quality of life of the people and the government’s concern for their citizen’s security. In addition, population size is compared to the necessity for a universal healthcare system; the lower the population, the more the advantages of digital transformation are realized. As a result, secure internet servers, UHC service coverage index, and population are the three most essential indicators in this case study.

4.3. Health Service Digital Transformation Benchmarking in ASEAN

Based on the absolute weight obtained in the above section, the relative weight of indicators is determined according to Equation (18). Then, the level of dominance $\left({\theta }_{ik}\right)$ matrix is constructed as Equations (19) and (20) with the loss attenuation coefficient equal to 60. This matrix is presented as

Table A6. The level of dominance with the loss attenuation coefficient $\mu =60$.

Country	BN	KH	ID	LA	MY	MM	PH	SG	TH	VN
BN	0.000	1.356	0.697	1.866	0.598	1.240	0.741	−0.241	0.170	0.453
KH	0.684	0.000	0.697	1.811	0.945	1.035	0.850	0.119	0.479	0.393
ID	0.414	1.108	0.000	1.843	0.360	1.132	0.751	−0.155	−0.076	0.420
LA	−0.196	−0.153	−0.285	0.000	−0.280	0.007	−0.274	−0.351	−0.349	−0.323
MY	0.568	1.176	0.524	1.686	0.000	1.404	0.700	−0.036	0.137	0.663
MM	0.293	0.444	0.520	1.433	0.499	0.000	0.569	−0.175	−0.122	−0.215
PH	0.379	0.807	0.304	1.827	0.385	1.155	0.000	0.002	−0.020	0.123
SG	1.800	2.097	1.900	2.730	1.860	2.249	2.120	0.000	1.445	1.542
TH	1.239	1.526	1.267	2.468	1.120	1.714	1.417	0.035	0.000	0.812
VN	0.797	1.247	0.932	2.347	0.856	1.529	1.080	0.018	0.300	0.000

in Appendix. According to Equation (21), the total dominance ${\theta }_i^{\ast }$ of ASEAN countries is calculated and illustrated in Figure 3. Ultimately, the overall evaluation score of countries is determined. As shown in Figure 4, one can clearly see that Singapore, Thailand, and Vietnam are leading the digital transformation for healthcare services in ASEAN. Singapore’s lead is relatively understandable given the maturity of technology and infrastructure as well as its low population requirements. In contrast, with a much higher population, the successes of Thailand and Vietnam in healthcare digital transformation are significant benchmarking for ASEAN.

5. Sensitivity and Comparison Analysis

5.1. Sensitivity Analysis

5.1.1. Indicators’ Weights

To evaluate the sensitivity of the indicators’ weights to the benchmarking results, firstly, the indicators are classified into three groups as shown in

Table 7. Indicator classification.

Group	C1	C2	C3	C4	C5	C6	C7	C8
Technology-related	X	X		X			X
Governance-related			X			X
Human-related					X			X

. Then, different weighting scenarios were suggested as shown in Figure 5. The scenarios 1, 2, and 3 focus on the indicators in technology-related, governance-related, and human-related groups, respectively. The benchmarking results are recalculated for the scenarios with a fixed loss attenuation coefficient $(\mu =60)$ as illustrated in Figure 6.

Figure 5 indicates that the best-performing nations are Singapore, Thailand, and Vietnam, while the lowest-performing countries are the Philippines, Myanmar, and Laos, regardless of the weighting situations. The rankings of the remaining nations have changed. Cambodia is placed fourth in the baseline scenario and seventh in scenarios 1 and 2. This demonstrates that a focus on criteria relating to technology and government management would boost Cambodia’s standing within the ASEAN region.

5.1.2. Loss Attenuation Coefficient

The loss attenuation coefficient is first divided into six points, as shown in Figure 7 to test the relevance of the weights of the indicators to the benchmarking results. Moreover, regardless of the loss attenuation coefficient, Figure 7 reveals that Singapore, Thailand, and Vietnam have the highest performance, while Myanmar and Laos have the lowest. The remaining countries’ standings have altered. Cambodia ranks eighth for coefficients less than or equal to 12.39 and fourth for coefficients equal to 51.47. This suggests that when the coefficient is raised, decision makers are not too concerned with the possibility of boosting Cambodia’s status in the ASEAN area.

5.2. Comparison with Other MCDM Method

For validation, in this section, the ranking results are determined based on the Final Utility matrix by TOPSIS and EDAS methods. Figure 8 illustrates the ranking results of the proposed method with loss attenuation coefficient equal to 60, TOPSIS, and the Evaluation Based on Distance from Average Solution (EDAS) method. The comparison shows that there is no significant difference in ranking results with the same weights of criteria. Accordingly, the Philippines has the greatest fluctuations in ranking. However, the proposal method allows adjusting the psychological behavior coefficients to better suit the decision maker.

6. Conclusions

In recent years, techniques for determining the objective weight of the criteria, which are based on the nature of data through statistical parameters, are researched and developed rapidly. In addition, psycho-behavioral theories are increasingly being applied in bounded rational decision-making problems. Motivated by these trends, this study developed and introduced a novel MADM approach, Psychological Behavioral Distance-based Decision-making Method. In the proposed approach, the weights of the criteria or indicators are determined through their distance-based correlation in the decision matrix. Furthermore, the principles of prospective and discrete theory are also applied in the computational procedure in the form of coefficients to appropriately express the psychological behavior of the decision maker. The proposed approach is then applied to the healthcare digital transformation benchmarking in ASEAN. The results show that the weighting of the indicators can be objectively determined by the proposed method. Singapore, Thailand and Vietnam are identified as the leading countries in healthcare digitalization in ASEAN. Moreover, through comparison with other MADM methods as well as analysis of sensitivity, validity and flexibility, the proposed method is evaluated.

Regarding the theoretical contribution, which is also the primary contribution, this study has introduced a novel approach that both ensures the objectivity of the data and meets the flexibility according to the subjective attitude of the decision maker. Regarding practical contributions, the study provides valuable insights into the factors that influence digital transformation in healthcare services in ASEAN countries. The findings suggest that infrastructure quality, information and communication technologies, and healthcare system responsiveness are crucial factors. The study also provides a comprehensive evaluation of the performance of different countries and identifies the leading countries in digital transformation in healthcare services in ASEAN.

Future research could further examine this technique for other psychological behaviors, such as risk, regret, and loss aversion, and consider more advanced aspects that affect the decision maker. Sensitivity analyses for all coefficients could also be evaluated to emphasize their influence in certain scenarios. Additionally, a new method could be developed and applied for coefficient testing. Other MCDM methods could be compared with this proposed approach, and the outcomes of this study could be a helpful resource for academics or investors seeking to implement this novel method in other sectors.